RE: [mpich-discuss] mpich2-1.0.7 file mpif.h and fortran compiler g95

[cornelis.broeders at web.de]

Using a parameter like "-DG95" or e.g.  "-DNO_SAVE_MPICH2" during configure seems to be no problem. The use of "blanket save" in a code is responsability of the code developer. However, best solution seems to be that the g95 compiler gives a "warning" instead of an "error" in this case. Do mpich2 developer have contacts to g95 developer? This would be the simplest way to solve the problem.
C.

> -----Ursprüngliche Nachricht-----
> Von: "Rajeev Thakur" <thakur at mcs.anl.gov>
> Gesendet: 14.10.08 04:40:28
> An: <mpich-discuss at mcs.anl.gov>
> Betreff: RE: [mpich-discuss] mpich2-1.0.7 file mpif.h and fortran compiler g95


> > I think this argument indicates that my proposal is not 
> > generally usable. On the other hand, the type of coding of 
> > mcnpx (reduced to a simple example in one of my mails) is 
> > quite convenient and works with most compiler. I am not 
> > familiar with typical including of "special cases" in mpich2 
> > coding strategy, but would it be possible to control the 
> > "SAVE" statements by a "configure" parameter to avoid "manual 
> > patching" of mpih.f in this special case?
> > C. 
> 
> One would need access to the user's code at configure time. And what if the
> user had many codes, some with blanket save and some without.
> 
> Rajeev
> 
> 

-- 
http://www.cornelis-broeders.de 




C.H.M. Broeders,       http://www.cornelis-broeders.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Cornelis Broeders.vcf
Type: text/x-vcard
Size: 279 bytes
Desc: not available
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20081014/11dd2da7/attachment.vcf>